Atmospheric Dispersion Modeling
Professor Tim LarsonCEE 357
Types of Models• Source Receptor
– Dispersion Calculations– Wind Tunnel– Empirical Scaling– Linear Rollback– Non-linear (chemistry/deposition)
• Receptor Source– Deduce “source fingerprints” (statistical)– Microscopy (particle shape & composition)
• Receptor Receptor– Forecasting and interpolation – Spatial and temporal
Dilution vs Dispersion:The Importance of Averaging Time
The instantaneous puffs are dispersed by fluctuating wind directions.
Instantaneous Conc. Profile
Time Averaged
Profile
“Puffs” of Pollution
Wind
The instantaneous concentrations are not described by steady-state plume models. The time averaged values are.
Dispersion models describe time-average plumes
Source: Slade et al “Meteorology and Atomic Energy, 1968”
instantaneous time-averaged
InstantaneousPlume Shape
Time-averaged Plume Shape
Heating of the surface
Unstable Daytime Conditions
Sunny, clear skiescolder, denser air
warmer, lighter air
Visible radiation to groundLarge scale vertical motions
Cooling of the surface
Stable Nighttime Conditions
Clear skies
colder, denser air
warmer, lighter air
Infrared radiation to space Vertical motions suppressed by density gradient
Neutral Conditions (day or night)
relatively high wind speeds
Motions not affected by buoyancy forces
Vertical air motions due to shear stresses
Minimal heating or cooling of the surface
Pasquill-Gifford-Turner Stability Classifications
(view from side)
Strongly Unstable
(view from side)
Stable
Surface Level Impacts Varywith Meteorology and Release
Height
high/moderatesurfacemoderate/lowelevatedneutralvery highsurfacenot significantelevatedstablehigh/moderatesurfacehighelevatedunstableSurface ImpactsRelease HeightStability
Fig 4-3, p.44 in Martin et al
wind
Describing Plume Concentrations
Factors Affecting Atmospheric Dilution (Mixing)
• Wind Speed
Concentration is inversely proportional to wind speed
Tim’s Simple Plume Model
h
x
12
3
mass/timepassingpoint 1
=mass/timepassing thrudisk area 2
mass/timepassing thrudisk area 3
C1 > C2 > C3
=
Simplified Steady-State Plume Model*Pollutant is well mixed and confined within the cone*Pollutant is continuously swept thru the cone by the wind
Concentration vs. distance downwind depends upon cone shape
2disk of areaspeed windrateemission Mass 2at air of Conc.
1
2
Simple Model #1:
23 mm/secg/sec
mg
Disk shape depends upon stability category
More unstable and thus more pronounced vertical spreading
unstable
neutral
stable
Heating of the surface
Sunny, clear skiescolder, denser air
warmer, lighter air
Visible radiation to groundLarge scale vertical motions
Cooling of the surface
Clear skies
colder, denser air
warmer, lighter air
Infrared radiation to space Vertical motions suppressed by density gradient
AA
Time-averaged concentration across AA
Gaussian (normal) distribution occurs across AA due to changes in wind direction over averaging period
Shape is described by “plume sigmas”
Most probable wind direction
source
windMass is not uniformly distributed within the cone’s volume
More Detailed Plume Model
functionon distributiGaussian 2disk of areaspeed wind
rateemission Mass 2at Conc
1
2
Simple Model #2:
x
z
y
X is the time-averaged wind direction,Y is the cross-wind direction,Z is the vertical dimension
23 mm/sec
g/secmg
Gaussian Plume(Concentrations vary with x, y and z)
y and z are functions of x
PlumeCenterline
x
h = hs + h
Physical stack height,hs
Plume rise, h“Effective”
stack height, h
Plume “Reflection” off of the Ground(pollutant cannot penetrate the ground)
Actual Source
“Virtual” Source(below the surface)
Most of plume above the surface
Plume begins to “reflect” off the
surface
Reflected material (shaded area)
Resulting in an asymmetric
vertical profile
Reflection is modeled by adding a “virtual” source contribution to the “real” one
h
Gaussian “Point” Source Plume Model:
Wind speed evaluated at “effective” stack height
Mass emission rate
}
Corresponds to disk area in simple model (values depend upon stability class & downwind distance, x)
Distribution of mass in vertical dimension (z) at a given downwind distance, x (includes the effect of surface reflection)
Distribution of mass in cross-wind dimension (y) at a given downwind distance, x
Pollutant concentration as a function of downwind position (x,y,z)
2
y
2
2
2
2
2
zy 2σyexp
2exp
2exp
u 2Q z)y,C(x,
zz
hzhz
“Effective” stack height, including rise of the hot plume near the source
Gaussian Plume(Concentrations vary with x, y and z)
For a given x, the max conc. is at the plume centerline and decreases exponentially away from the centerline at a rate dependent upon the sigma values, y and z.
y and z are functions of x
PlumeCenterline
Cross-wind distance from plume centerline (m)
Verti
cal d
ista
nce
from
plu
me
cent
erlin
e (m
)
Concentration distribution in a Gaussian plume (y = 20 m; z = 10m; centerline concentration = 1.0)
Source: Hanna et al, 1981
Note: theoretical plume has infinite extent in all directions!
y
z
Calculation Procedure
1. Determine stability class from meteorology
2. Compute wind speed at “effective” stack height, h
3. Compute y and z at a given downwind distance, x
4. Choose appropriate receptor height, z5. Compute C(x,y,z) using Gaussian plume
equation
Pasquill-Gifford-Turner Stability Classifications
Calculation Procedure
1. Determine stability class from meteorology
2. Compute wind speed at “effective” stack height, h
3. Compute y and z at a given downwind distance, x
4. Choose appropriate receptor height, z5. Compute C(x,y,z) using Gaussian plume
equation
Calculating Wind Speed as a Function of Height
"Power Law" Method
This approach is used with the EPA models and employs a simple "power law" function. The wind speed at any elevation is estimated as a function of the height of the actual wind speed measurement, the stability category, and the "wind profile exponent", as follows:
Uz = Uzrefz
zrefp
Whereuz = wind speed at some arbitrary elevation z [meters] uzref = wind speed at the "reference" (actual measurement) height [m/sec]zref = the elevation of the actual wind speed measurement [m] p = wind profile exponent, a function of stability category.
Values of p as a function of stability category are summarized in the following table. These are the “default” p values recommended by EPA for use when zref = 10 m.
Stability Category Urban RuralA 0.15 0.07B 0.15 0.07C 0.20 0.10D 0.25 0.15E 0.30 0.35F 0.30 0.55
The "urban" and "rural" classifications attempt to capture the effect of surface roughness. The largest effect is seen under very stable conditions (“F”).
Example Calculation:
For the “rural’ case, if the wind speed is 3 m/sec measured at an elevation of 10 meters and the stability category is "D", then p = 0.15 and the wind speed at z=100 m is:
U100 = 3.5 m/s 100 m10 m
0.15 = 4.94 m/s
There are other ways to estimate the wind speeds as a function of height, but the “power law” approach is probably the simplest and most straightforward method.
Calculation Procedure
1. Determine stability class from meteorology
2. Compute wind speed at “effective” stack height, h
3. Compute y and z at a given downwind distance, x
4. Choose appropriate receptor height, z5. Compute C(x,y,z) using Gaussian plume
equation
Sigma-y
x
Sigma-z
tan11628.465 xy
xdc ln017453293.0
x is in kilometersy is in meters is in radians
Cross-wind distribution:
bz axVertical distribution:
x is in kilometersz is in metersa, b depend on x
Plume sigma formulas from EPA’s ISC Model
Pasquill Stability Category
x (km)
a b
A* <.10
0.10 - 0.15
0.16 - 0.20
0.21 - 0.25
0.26 - 0.30
0.31 - 0.40
0.41 - 0.50
0.51 - 3.11
>3.11
122.800
158.080
170.220
179.520
217.410
258.890
346.750
453.850
**
0.94470
1.05420
1.09320
1.12620
1.26440
1.40940
1.72830
2.11660
**
* If the calculated value of σz exceed 5000 m, σz is set to 5000 m.
bz ax
B* <.20
0.21 - 0.40
>0.40
90.673
98.483
109.300
0.93198
0.98332
1.09710
C* All 61.141 0.91465
D <.30
0.31 - 1.00
1.01 - 3.00
3.01 - 10.00
10.01 - 30.00
>30.00
34.459
32.093
32.093
33.504
36.650
44.053
0.86974
0.81066
0.64403
0.60486
0.56589
0.51179
* If the calculated value of σz exceed 5000 m, σz is set to 5000 m. ** σz is equal to 5000 m.
Pasquill Stability Category
x (km)
a b
bz ax
Pasquill Stability Category
x (km)
a b
E <.10
0.10 - 0.30
0.31 - 1.00
1.01 - 2.00
2.01 - 4.00
4.01 - 10.00
10.01 - 20.00
20.01 - 40.00
>40.00
24.260
23.331
21.628
21.628
22.534
24.703
26.970
35.420
47.618
0.83660
0.81956
0.75660
0.63077
0.57154
0.50527
0.46713
0.37615
0.29592
F <.20
0.21 - 0.70
0.71 - 1.00
1.01 - 2.00
2.01 - 3.00
3.01 - 7.00
7.01 - 15.00
15.01 - 30.00
30.01 - 60.00
>60.00
15.209
14.457
13.953
13.953
14.823
16.187
17.836
22.651
27.074
34.219
0.81558
0.78407
0.68465
0.63227
0.54503
0.46490
0.41507
0.32681
0.27436
0.21716
bz ax
Pasquill Stability Category
c d
A 24.1670 2.5334
B 18.3330 1.8096
C 12.5000 1.0857
D 8.3330 0.72382
E 6.2500 0.54287
F 4.1667 0.36191
xdc ln017453293.0
Calculation Procedure
1. Determine stability class from meteorology
2. Compute wind speed at “effective” stack height, h
3. Compute y and z at a given downwind distance, x
4. Choose appropriate receptor height, z5. Compute C(x,y,z) using Gaussian plume
equation
Steady State Gaussian “Point” Source Plume Model:
Wind speed evaluated at “effective” release height
Mass emission rate
}
Corresponds to disk area in simple model (values depend upon downwind distance, x)
Distribution of mass in vertical dimension (z) at a given downwind distance, x (includes the effect of surface reflection)
Distribution of mass in cross-wind dimension (y) at a given downwind distance, x
Pollutant concentration as a function of downwind position (x,y,z)
2
y
2
2
2
2
2
zy 2σyexp
2exp
2exp
u 2Q z)y,C(x,
zz
hzhz
“Effective” stack height, including rise of the hot plume near the source
Example Calculation
2
y
2
2
2
2
2
zy 2σy-exp
2exp
2exp
u 2Q z)y,C(x,
zz
hzhz
Given:Q = 10 grams/sec; h (=heff) = 50m; x = 500 m = 0.5 km; u50 = 6 m/s; Stability Class “D”Compute:C(500, 0, 0) ,i.e., the ground level concentration (z = 0) at plume centerline, 500 meters downwind.
18.3m)32.093(0.5axσ 0.81066bz
radians1542.0)0.50.72382ln-(8.3330.0174532930 m1.36)1542.0tan()5.0(11628.654)(tan11628.654σy x
2
2
2
2
2
2
36.120-exp
3.182500
exp3.182500exp
3.181.366 210 C(500,0,0)
335 g/m19.2g/mx1092.110.04793.181.366 2
10 C(500,0,0)
2
y
2
2
2
2
2
zy 2σy-exp
2exp
2exp
u 2Q z)y,C(x,
zz
hzhz
2
y
2
2
2
2
2
zy 2σy-exp
2exp
2exp
u 2Q z)y,C(x,
zz
hh
2
y
2
2
2
zy 2σy-exp
2exp 2
u 2Q z)y,C(x,
z
h
Simplified Plume Equation for z = 0
x
Another classic computation involves finding the location and value of the maximum ground level concentration downwind of a tall stack.
Physical stack height,hs
Plume rise, h“Effective”
stack height, h
Concentration vs. x at ground level (z=0)
(note maximum at x > 0)
Concentration in elevated plume at height z
Non-Gaussian Plumes
Non- Gaussian Plume “Trapped” in Building Wake
Non-Gaussian Plume “Looping” During Unstable Conditions(large-scale vertical motions)
ExtremeDeparture
From Gaussian
Large-Scale Vertical Motions
The newest EPA plume model is called ‘Aermod’ and incorporates this downdraft effect.
60% of the time40% of the time
Time-averaged Plume is skewed downward
w* = vertical scaling velocity
Aermod uses two superimposed Gaussian models, one for downdrafts and one for updrafts
Water (stable)
Land (unstable)
source
‘Fumigation’
Plume Fumigation During On-shore Flow
wind